Influence of solvent on the magnitude of the anomeric effect J.-P. PRALY' AND R. U. LEMIEUX' University of Alberta, Edmonton, Alta., Canada T6G 2G2 Received July 3, 1986

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

This paper is dedicated to Dr. 0.E. (Ted) Edwards Can. J. Chem. 65, 213 (1987). J.-P. PRALY and R. U . LEMIEUX. A novel application of I3cnuclear magnetic resonance provided the effects of solvent polarity and hydrogen-bond formation on the conformational equilibria for a range of 2-substituted tetrahydropyrans and the results are interpreted in terms of how solvent affects the competition between the endo- and exo-anomeric effects in determining the magnitude of the anomeric effect. In accord with the generally accepted origin of the endo- and exo-anomeric effects (anti-periplanar n-a* interaction of the oxygen lone-pair orbital with the antibonding orbital of the adjacent C-0 bond), the exo-anomeric effect for the a anomer is expected to be weaker because charge delocalization from the glycosidic oxygen to anomeric center is in competition with delocalization from the ring-oxygen atom. The effects of solvent on the relative magnitudes of the endo- and exo-anomeric effects are then considered to arise from the formation of specific complexes with the solvent, and the exo-anomeric effect of a P-glycoside is more strongly influenced. It is contended that hydrogen bonding of solvent to the ring oxygen increases the exo-anomeric effects. For this reason water is particularly effective for the strengthening of the exo-anomeric effect and, thereby, the conformational rigidity of glycosides. Experimental evidence is presented that indicates that the anomeric hydroxyl groups of free sugars dissolved in water tend to prefer the equatorial orientation because these provide stronger hydrogen bonds as proton donors to water. Can. J. Chem. 65, 213 (1987). J.-P. PRALY et R. U. LEMIEUX. Une nouvelle application de la rmn du I3cpermet de dtterminer les effets delapolaritt du solvant et de la formation de liaisons hydrogknes sur l'tquilibre conformationnel d'une strie de tttrahydropyrannes portant un substituant en position 2; on interprkte les rtsultats en fonction de comment les solvants affectent la compttition entre les effets anombres endo et exo dans la dktermination de l'amplitude de l'effet anomkre. En accord avec l'origine gtntralement acceptte des effets anomkres endo et exo (interaction n-a* anti-pkriplanaire de l'orbitale de la paire libre de l'oxygkne avec l'orbitale antiliante de la liaison C-0 adjacente), il est prtvu que l'effet anomkre-exo de l'anomkre-a sera plus faible p a c e que la dtlocalisation de la charge de l'oxygkne glycosidique vers le centre anomkre est en compttition avec la dtlocalisation de l'atome d'oxygkne du cycle. On considkre alors que les effets de solvant sur les amplitudes relatives des effets anomeres endo et exo proviennent de la formation de complexes sptcifiques avec le solvant et que l'effet anomkre-exo d'un P-glycoside est plus fortement influenct. On pretend que la liaison hydrogene du solvant avec l'oxygene du cycle augmente les effets anomkres-exo. Pour cette raison, l'eau est particulierement efficace pour rehausser l'effet anomere-exo et, finalement, la rigiditt conformationnelle des glycosides. On prtsente des donntes exp6rimentales qui indiquent que les groupements hydroxyles anomeres de sucres libres dissous dans l'eau tendent ?I prtferer l'orientation tquatoriale parce que celle-ci donne lieu ?I des liaisons hydrogenes plus fortes comme donneurs de protons Sl l'eau. [Traduit par la revue]

II I

I

Introduction

I

The term "anomeric effect" was introduced in 1959 (1) to indicate the existence of a stereoelectronic influence on chemical bonding that had been detected (2) in a study of the anomerization equilibria for acetylated glycopyranoses of known ring conformation (3). Since then, the subject has received major attention (4, 5). It was initially proposed (6) that the anomeric effect was the result of bond dipole - bond dipole interactions. Mainly o n the basis of X-ray crystallographic data, Altona (7) suggested that the anomeric effect was a manifestation of charge delocalization (no-bond resonance) from the oxygen atom of the acetal group toward the anomeric carbon. Thus, the stabilization gained was greater for the anomer with the aglycon in axial orientation since this placed the polar glycosidic bond anti-periplanar to an unshared pair of electrons of the ring oxygen. The subject then drew very substantial attention from theoretical chemists. Wolfe et a l . (8) have presented a fine review of these developments. The nature of the orbitals and how these are considered to interact for the

I I

I

1

'university of Alberta Postdoctoral Fellow, 1983-1985. Present address: Universite Claude-Bernard Lyon 1, Ecole Superieure de chimie ~ ~ d ~de L ~ ~~ ~i ~~ ~l ~l de , ~Chimie b Organique ~ 2, ESCIL, 43 Boulevard du ll-Novembre-1918, 69622 villeurbanne CEDEX, France. 2 ~ u t h oto r whom correspondence may be addressed.

stabilization of a n acetal bond has been qualitatively presented in detail by Kirby (4). The differences in conformational energy for the various conformers for dimethoxymethane as estimated by the ab initio molecular orbital calculations of Jeffrey, Pople, a n d co-workers (9) were used in this laboratory (10) to assign a value to what had been termed the exo-anomeric effect (1 1). Subsequently (12), this contribution to the energy of a glycosidic bond was expressed in terms of two torsion expressions, namely [l]

exo-Ap = 2.61(1

- cos 8")

- 1.21(1 - cos 28")

- 1.18(1 - cos 38")

[2]

+ 2 . 8 6 kcal/mol

exo-A, = 1.58(1 - cos 8") - 0.74(1 - cos 28") - 0.70(1

- cos 38")

+ 1 . 7 2 kcal/mol

That the exo-anomeric effect appears some 1.66 times stronger for the anomers (1) that have the aglycon i n equatorial orientation (normally the P anomers) is not surprising, even in the absence of theoretical calculation. A s seen in formula 1 , both the unshared pairs of electrons of the ring oxygen are syn-clinal to the glycosidic bond and, therefore, unfavorably disposed for toward carbon. Thus, ~ ~ delocalization ~ ~ the anomeric i ~ ~ there exists n o important competition for delocalization of charge (back bonding) (13) from the glycosidic oxygen that, in the sterically most favorable orientation, has an unshared pair

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

214

CAN. I. CHEM. VOL. 65. 1987

of electrons anti-periplanar to the C1-ring oxygen bond. In contrast to the p anomers, the a-glycosides have the aglycon in axial orientation and therefore, as displayed in 2, are importantly stabilized by an endo-anomeric effect. However, for these a-anomers (2), competition exists between the endo and exo-anomeric effects for the electron deficiency at the anomeric carbon. Consequently, in accord with calculation (9, 14), the exo-anomeric effect for 2 should be weaker than that for the p anomer (1). Fuchs et al. (15) examined the X-ray crystallographic data for I I I carbohydrate structures and found only orientations for the aglycon that are compatible with stabilization by way of an exo-anomeric effect. The values for the exo-anomeric effects provided by expressions [I] and [2] are of course only a hopefully useful approximation. Certainly, the values for these stereoelectronic contributions to molecular stability will depend on the electronegativities of the substituents, both on the pyranose ring and on the aglycon (16). As will be discussed below, the solvent also appears to influence the values of exo-anomeric effects. Nevertheless, it was considered possible that the values for the a and p exo-anomeric effects, in conformers for dimethoxymethane as derived by ab initio molecular orbital calculation (9), would serve usefully for the molecular modelling of oligosaccharides. To date, this has proven to be the case. It was found, for example, that the conformational preference required by 'H nrnr data for the B human blood group trisaccharide determined in aqueous solution was not provided by hardsphere calculation (12). However, addition of a contribution by the exo-anomeric effect to the hard-sphere calculations (12) provided a structure that was not only in accord with the nrnr data but also present in the crystal structure (17). Despite much such definitive evidence for the importance of the exo-anomeric effect (12, 13), this stereoelectronic feature continues to be disregarded in efforts to assess the conformational preferences of oligosaccharides. For example, Lipkind, Verovsky, and Kochetkov (18) recently published on the conformational states of cellulose, lactose, and maltose, and state, in the absence of acceptable experimental support, that the exo-anomeric effect does "not play a significant role under the conditions of an aqueous medium." In fact, as will be demonstrated below, the exo-anomeric effect appears to be not only of major importance but even of special importance in aqueous solution. The main purpose of this communication is to provide experimental support for this contention. Indeed, the neglect of a consideration of the exo-anomeric effect renders obsolete a large literature on the theoretical calculation of the conformational preferences of glycosidic structures. The fact that the exo-anomeric effect adds rigidity about the glycosidic bonds of oligosaccharides is a matter of major importance to an understanding of their interactions with the receptor sites of proteins to form specific carbohydrate-protein complexes. In this regard, it has become apparent that oligosaccharides present well-defined and substantially rigid surfaces about which the protein becomes organized (19). The conformational analyses of compounds that contain polar bonds and atoms with unshared pairs of electrons require a

consideration of specific solvent effects. For example, the conformational equilibrium for methyl 3-deoxy-P-L-erythropentopyranoside was more sensitive to the chemical nature of the solvent than to its polarity (20). Similarly, specific solvation effects were found to influence the conformational equilibrium for 2-methoxytetrahydropyran (2-MeO-THP) (3) (1 1). The equilibrium data presented in Table 1 confirm the observation made in 1969 (1 1) that the polar solvent CD3CN favors the more polar equatorial conformer 3 e to a greater extent ?Me

3a

3e than does the nonpolar CC14. In D20 the two conformers have a near equal population (1 1). Thus, the anomeric effect for 3 is weak in water. Also, the mutarotation of many aldopyranoses in water favors the anomer with the anomeric hydroxyl in equatorial orientation (4). It is these kinds of observations that have led to the widespread notion that the anomeric effect is unimportant when the solvent is water. However, the stereoelectronic factors that are responsible for the anomeric effect cannot be expected to simply disappear in certain solvents. Instead, it must be considered that stereoelectronic effects on the strengths of the two C-0 bonds of acetals remain but are exhibited differently in different environments because of changes in the interactions involving the unshared pairs of electrons. That this is the case clearly evident from the fact that, as seen in Table 1, weakly polar CDC13 is as effective as CD3CN in favoring the more polar (1 1) conformer 3e. In contrast, Tvaroska and Kozar (21) have presented a theoretical study of the hydration of the acetal segment in glycosides and concluded that hydration does not operate against the anomeric and exo-anomeric effects. Also, as already mentioned, Lipkind et al. (18) concluded, on the basis of a theoretical conformational analysis of disaccharide derivatives in aqueous solution, that the exo-anomeric effect and intramolecular hydrogen bonding are of no significance. We disagree. It is to be kept in mind, however, now that the origin of the anomeric effect has become established (8, 9), that an anomeric effect (AA) must be considered as the difference between the sum (A3,) of the endoand exo-anomeric effects in the equatorial conformer and the sum (A3a) of these for the other conformer (22, 23). It is on this basis that the anomeric effect for 3 is defined as shown in expression [3]. As was mentioned above, the endo-A, contributions are considered negligible and, therefore, Ak is assumed to be equal to exo-A3,.

On the basis of eq. [3], anomeric effects will have a wide range of values, including negative and positive values, depending on the relative magnitudes of the various exo- and 3e, endo-anomeric effects. Indeed, for the equilibrium 3a the axial conformer 3a is strongly favored in CC14 and, therefore, A3emust be less negative (less stabilizing) than AJa in order that AA3 be positive. In contrast, 2-methylaminotetrahydropyran (4) because of the basic nitrogen is expected to possess strong exo-anomeric effects. The exo-A4, effect should be greater than the exo-A4, contribution to the stability and, perhaps, sufficiently greater for the equilibrium to favor 4e even though 4a should be additionally stabilized by an endo-

PRALY AND

TABLE 1. The conformationalequilibria for 2-methoxytetrahydropyran in a variety of solvents at 308 K as estimated from studies of its 4,4,5,5-tetradeuterio derivative (6) Mole fraction (n,) of the equatorial conformer Present study

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

Solvent

cc14

I

c6D6 CCI4-C6D6 (10%) cc14-c6F6 ( 1 0%) CDC13 (CD3)zCO (CDdzSO CD3CN DzO

Dielectric constant Lit. (1 1) 2.2 2.3 -2.2 -2.3 4.8 21 36 38 78

0.17 0.18 -

0.29 0.28 0.26 0.32 0.48

JH,Ha nOeb -

-

0.20 0.20 0.33

0.18 -

0.34

-

-

0.37 0.48

0.37

-

JHDC

0.21 0.26 0.36 0.37 0.47

"From the time-averaged coupling of H-2 with H-3c and H-3t. bFrom the relative enhancements (R) of the signals for H-6c and the methyl group on saturation of H-2 of 6 . Setting R = nOe (H-6c)/nOe (CH3), R = 0.18 (CC14-C6D6), R = 0.34 (CDCI,), R = 0.37 (CD3CN). 'From the coupling between H-6c and H-6t and deutenum atoms at C-5.

anomeric effect. In fact, Booth et al. (24) havejecently shown 4e favors the equatorial conformer. that the process 4a De Hoog (25) had earlier demonstrated a similar situation for 2-dimethylaminotetrahydropyran (5). On the other hand, the

+

I

:

substitution of the ring oxygen of a glycosidic structure by the stronger electron-donating NH group should increase the endo-anomeric effect of an a anomer but decrease the strengths of the exo-anomeric effects. Indeed, Pinto and Wolfe (26) have shown that the antibiotic Nojirimycin (5-amino-5-deoxy-Dglucopyranose) mutarotates in water to an equilibrium a / P = 1.7, in contrast to D-glucopyranose for which a / P = 0.56. The conformational analysis of the equilibrium 3a 3e can be expressed as

+

;

I 1

I

where AAO = anomeric effect, AR0 = change in nonbonded interactions, and AS0 = change in entropy. AGO is positive since 3a is favored. For compound 3, in nonpolar solvents, ASo should be negligible. The highest value to date for the anomeric effect (AAO) is the 2.1 kcal/mol proposed by Franck (27) using CC14 as solvent. The change in nonbonded destabilizing interactions (AR0) is not known but, becauseof the short C-0 bond in the tetrahydropyranyl ring, should be greater than the A value (0.6 kcal/mol) (28) for a methoxy group. Assuming that 3a is destabilized by 1.1 kcal/mol and that the change in entrop is negligible (27), the change in enthalpy (AHO = AA0 + AR ) would appear to be about 1 kcal/mol in favor of 3a. That this result has some validity is indicated by the fact that, with CCl, as solvent, AGO is near 1 kcal/mol (1 1) (seeTable 5). However, Booth et al. (24) found AH for 3a 3e with CFC13-CDC13 as solvent to be near zero. It was suggested that

8

+

3a may be favored (K, = 0.256) because of an unusually high difference in conformational entropy. On the other hand, Lemieux et al. (1 1) had provided evidence that 3 forms specific complexes with polar solvents and especially with solvents such as CDC13 and D 2 0 , which can provide a proton to hydrogen bond formation. Thus, it seemed possible that the near temperature independence of the change in free energy noted by Booth et al. (24) was due to stronger bonding of the solvent (CFC13/CDC13) with 3e to form solvated species of nearly equal enthalpy contents. Since CDC13 solvent molecules are less prone to self-association than to bonding with the acetal (3), the decrease in entropy for the solvation of 3e should be greater than that for 3a. Thus, the net change in entropy should favor 3a, as was observed. An examination of this possibility was a main purpose for this investigation. Deslongchamps (5a) has discussed the im~ortanceof stereoelectronic factors to the reactivity of acetals and, in this regard, pointed out (5b) that stereoelectronic effects should "influence the basicity of the oxygen atoms of the acetal function, thence their relative ease of protonation." The results to be presented below are in strong support of this contention. Discussion of results As already mentioned, this investigation was undertaken mainly because the conformational equilibrium for 3 in CFC13CDC13 (85: 15 v/v) was found to be essentially independent of temperature (23). The extension of temperature studies to other solvents appeared possible by the procedure used in 1969 (1 l), which involved measurement of the time-averaged values of J2,3& and J2,3trans for 4,4,5,5-tetradeuterio-3; namely, the equilibrium 6a S 6e.

The conformational equilibria for the tetradeuterio compound 6 were estimated using the same procedure as previously described (1 1) except that the spacings in the 'H nmr spectra were more precisely determined by operating at 400 MHz with deuterium decoupling. The results obtained using five different solvents and two different temperatures are presented in Table 2. As was previously done ( l l ) , the J1,2a and J1,2e coupling constants for the methyl 2-deoxy-a- and Zdeoxy-P-D-arabinohexopyranosides were used to calculate the conformational equilibria from the observed values for J2.3c and J2.3r.These equilibria are presented in Table 1, where it is seen that the results at 308 K agree well with those previously obtained (1 1) using a spectrometer operating at 100 MHz in the CW mode. The use of the coupling constants found for the 2-deoxyglucosides to calculate conformational equilibria for 2-MeO-THP (3) is, of course, arbitrary. That the procedure provides results useful for the present purposes was previously (1 1) supported by the optical rotations of an optically active sample of 2-MeO-THP in the various solvents. This conclusion is now supported, as described in the Experimental and involving structures 7-10, by two other procedures: (a) the coupling of the H-6 atoms with deuterium atoms at C-5, and (b) the nuclear Overhauser enhancements (nOe) observed o n saturation on H-2. The above-described procedure for the determination of conformational equilibria of 6 by 'H nmr was prohibitively

216

CAN. J. CHEM. VOL. 65, 1987

TABLE2. 'H nuclear magnetic resonance data for the estimation of changes in the conformational equilibrium of 2-methoxytetrahydropyran (3) with changes in solvent and temperature Chemical shifts (ppm)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

Solvent

H-2

H-3c

H-3t

H-6c

H-6t

Coupling constants (Hz) 0CH3

H-2,H-3c

H-2,H-3t

H-3c,H-3t

H-6c,H-6t

n6e

H-2a,H-3

H-2e,H-3

2. Methyl 2-deoxy-D-arabino-hexopyranosidesin D20 at 308 K

a Anomer

p Anomer

H-1

H-2a

H-2e

H-3

H-4

0CH3

H-1,H-2a

H-l,H-2e

H-2a,H-2e

4.90 4.61

1.70 1.46

2.13 2.24

3.85 3.70

3.35 3.24

3.35 3.51

3.71 9.79

1.32 2.00

-13.51 - 12.38

expensive, both in time and facilities, for broader application. However, it proved possible using the mole-fraction data presented in Table 2 to establish an efficient procedure based in 3~ nmr for the study of the conformational equilibria, not only under a variety of solvent conditions but also for a range of different 2-substituted tetrahydropyrans. The procedure, which is described in the Experimental, involves the assignment of "absolute" chemical shifts for C-4 in the conformations 3a and 3e. On this basis, time-averaged chemical shifts for C-4 could be interpreted to yield the mole fractions of 3a and 3e under the prevailing conditions. This procedure is made possible by the great stability of the magnetic fields produced by the superconducting solenoids of the modem high-frequency nmr spectrometers. The equilibrium data at 308 K presented in Table 3, using Ccl4-C6D6 (10%) as solvent, are in accord with expectations based on the nature of the anomeric effect as it exists in the literature (4) and as discussed in the Introduction. The electronegativities of the aglycons CH2FCH2,CH20AcCH2, and CH3 are similar and, indeed, the conformational equilibria are all in the range 0.22 + 0.02. As expected, the monofluoro compound (12) most favors the conformer with the aglycon in axial

11.87 11.76

5.27 5.09

has long been appreciated (29), but is now considered to occur in part because the CF3CH20 group of 11 strengthens the endo-anomeric effect for conformer l l a to a greater extent than does the less electronegative CH2FCH20 group of 12. Also, however, the more electronegative CF3CH2 group of 11 weakens both the exo-anomeric effects to a greater extent than does the CH2FCH2 group. As already discussed with reference to the amino compounds 4 and 5, the equatorial orientation for the aglycon of a 2-substituted THP compound will be favored the stronger the exo-anomeric effect-an effect that increases with increasing ease for charge delocalization from the aglycon to the anomeric carbon. On this basis, the trimethylsilyl ether 14 was expected (30) to favor conformer 14e, as compared to 3 favoring 3e. In fact, the equilibrium for 14 in Ccl4-C6D6 is much more in favor of the equatorial conformer than is that for 3 and, for that matter, also those for 12 and 13.

.

P

Si Meg

C--

X

140

11, 12,

X=

13,

X=

X

CF~CH~O

=C H 2 ~ C ~ 2 0 CH20AcCH20

orientation. Replacement of the CH2F group of 12 by the much more electronegative CF3 group to form 11 is seen to greatly enrich the population of the axial conformer. This consequence

14e

As indicated in the conformational equilibrium for 14, the ring oxygen of the 14e conformer should serve best as proton acceptor-for hydrogen bond formation with a protic solvent. Therefore, the population of 14e at equilibrium should increase when the solvent is changed from carbon tetrachloride to chloroform (A = CC13). As seen in Table 3, the mole fraction of 14e does increase from 0.28 to 0.47 on changing the solvent from CC1,-C6D6 (10%) to CDC1,. In D 2 0 , the mole fraction of 14e was even greater at 0.70. This high abundance of 14e can be

PRALY AND LEMIEUX

TABLE3. Conformational equilibria for various 2-substituted tetrahydropyrans in different solvents at 308 K

Chemical shifts, ppm

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

2-Substituent

Observed C-4

attributed to both the high polarity of D20 as well as to its ability to behave as a "proton" donor for hydrogen bond formation. The polar CD3CN provided an equilibrium with nearly the same population of 14e as does the much less polar solvent CDC13 (1 1). In CC14-C6D6, the abundance of conformer 15e of the hydroxyethyl compound 15 is significantly greater (n15 = 0.3 1) than that (0.19) of conformer 3e for compound 3 although the CH20DCHz group of 15 is certainly more electronegative than the CH3 group of 3. This apparent anomaly can be attributed to involvement of the ring oxygen of 15 in hydrogen bonding with the hydroxyl of the aglycon as is displayed in the formulas for 15a and 15e. This driving force in favor of 15e

"Absolute" [c-41

AGO

nc

Kc

(kcal mol-I)

would be eliminated in solvents such as CD3CN and D20, which can act as good proton acceptors. In fact, the mole fractions of 15e in these solvents are greater and nearly the same as those for 12e, 13e, and 3e. On the other hand, CDC13is a very poor proton acceptor, and therefore 15e should be favored to a greater extent in this solvent than would be either 12e, 13e, or 3e. As seen in Table 3, this is the case and provides convincing evidence that engagement of the ring oxygen in hydrogen bond formation favors the equatorial orientation for the aglycon. Since the proposals made in the Introduction to rationalize the influences of solvent on the anomeric effect appear well supported, it was of interest to interpret the solvent effects on 2-hydroxy-THP (16) on the same basis. In fact, to do so not only provides an attractive rationalization of the effect of solvent on the conformational equilibrium of 16 but also on the anomerization equilibria for a number of sugars, such as glucose, which favor the equatorial orientation for the anomeric hydroxyl group. As seen in Table 3, the conformational equilibria for 16 in both the solvents CC14-C6D6 (10%) and CDC13 are nearly the

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

218

CAN. J. CHEM. VOL. 65, 1987

same as those for the hydroxyethyl compound (15). However, major differences in equilibria (0.58 versus 0.41 and 0.70 versus 0.50) occur when the solvent is CD3CN or D20, respectively. These differences are attributed to the fact that when dissolved in these latter solvents, both 15 and 16 become extensively hydrogen bonded as hydrogen donors to solvent molecules. The changes in bond polarity that arise from the formation of these hydrogen bonds should not much influence the electronegativity of the CH20HCH2 group of 15. In fact, 15e has nearly the same population (0.41) in CD3CN as in CDC13 (0.44). On the other hand, the population of 16e is significantly greater in CD3CN than in CDC13 (0.58 versus 0.49). This is as expected, since hydrogen bonding of the hydroxyl group of 16 with a solvent molecule must enrich the electron density on the oxygen atom (20) and thereby increase the exo-anomeric effects, but more so that for the equatorial conformer (16e). Thus, the proton-accepting solvent should, as was observed, enrich the population of the equatorial conformer (16e). The situation is similar with water as solvent. As seen in Table 3, the 0.50 mole fraction for 15e in D20 is nearly the same as those for 12e, 13e, and 3e, namely, 0.48, 0.48, and 0.47. On the other hand, the mole fraction for 16e is 0.70 in D20. This result is attributed to a greater strengthening of the exoanomeric effects, but more so for 16e as the result of the

~

~

I

hydrogen bonding with water as is indicated in the formula for 16e. Experimental evidence exists for increased electron density on an oxygen atom of an hydroxyl group as the result of hydrogen bond formation (20), and Jeffrey and Mitra (3 1) have provided evidence, based in crystal structures, that anomeric hydroxyls are in fact strong proton donors but poor proton acceptors for hydrogen bond formation. To better establish the influence of hydrogen bonding on the conformational equilibrium for 16, a study was made of the effect of the addition of dimethylsulfoxide to solutions in Ccl4-C6D6 (10%). The results, presented in Fig. 1, show that the population of 16e increases in a manner consistent with increasing hydrogen bonding of 16 with the DMSO, as displayed by the chemical shifts for both H-2 and OH-2. In contrast, the same addition of DMSO to solutions of 2-CH30-THP (3) had no effect on the chemical shift for H-2 and very little effect on the conformational equilibrium. We therefore abandon the earlier postulate that the influence of water on the conformational equilibria for 2-methoxytetrahydropyran (3) (11) and 2-hydroxytetrahydropyran (16) (32) is related to a more favorable water bridge between the two oxygen atoms when the conformer has the aglycon in equatorial orientation. In view of the above observations on the effects of solvents on the conformational equilibria for 16, we propose that certain sugars such as glucose, galactose, and xylose favor the p anomer when dissolved in water because of a strong P exo-anomeric effect. To further support this proposal, a brief study was made of the mutarotation of 2,3,4,6-tetra-0-methylD-glucopyranose(17) (33). The mole fraction of the P anomer, which has the anomeric hydroxyl in equatorial orientation, at

m 12 0.4

0.8

1.2

1.6

DMSO (molarity) in CCI4-C6D6(10%) FIG.1. The effects of additions of DMSO to 0.1 M solutions of 2-MeO-THP (A,8H-2;0, n,) and 2-HO-THP (0, 80Fl-2; v, 8Fl-2; n, n,) in CCI4-C6D6 (10%).at 308 K, on chemical shifts and the conformational equilibria.

equilibrium in water and at 298 K was found to be 0.39. This result contrasts with that for the parent D-glucose for which the mole fraction at equilibrium is 0.64. Evidently the situation is not as straightforward for these complex structures as it is for 16. Nevertheless, the mutarotation of the methylated glucose (17) (0.1 M) was examined in Ccl4-C6D6 (10%) and found to be well catalyzed at 298 K by the addition of 2-hydroxypyridone (34) (0.001 M). Under these conditions, at equilibrium, the mole fraction of the p anomer was 0.41. On rendering the solution 0.8 M in DMSO, the mole fraction of this anomer increased to 0.51, in accordance with the expectation that the hydrogen bonding with the DMSO more strongly strengthens the flexo-anomeric effect. It was earlier suggested, with reference to the conformational equilibrium for 2-hydroxyethyloxy-THP (IS), that the exoanomeric effect is enhanced by hydrogen bonding involving the ring oxygen (0-5). It was further proposed that this enhancement is greater for the equatorial conformer since, for this conformer, 0 - 5 is not importantly involved in an endoanomeric effect. Intramolecular hydrogen bonding is supported by the presence of a bond between the hydroxyl and the ring oxygen of 2(R)-hydroxy- 1(R)-cyclohexyl P -D-glucopyranoside tetraacetate in the crystalline form (35). To further test the hypothesis, the data reported in Table 4 were obtained. It is well known that phenol is a good proton donor to hydrogen bond formation. Therefore the addition of this substance to a solution of 3 in a nonpolar aprotic solvent such as CC14-C6D6 (10%) is expected to lead to hydrogen bonding between the phenol and the acetal group. As mentioned above, the ring oxygen of the equatorial conformer of 3 is expected to form the strongest hydrogen bond and, thereby, cause a shift of the conformational equilibrium toward a higher population of the equatorial conformer. As seen in Table 4, the addition of phenol at 308 K to a solution of 3 in CC14-C6D6 in fact led to an increase of n3, from 0.18 to 0.29. In contrast, a similar addition of

PRALY AND LEMIEUX

solvent than in the non-hydrogen-binding solvent Ccl4-C6D6 (10%). The results obtained with CD3CN and D 2 0 as solvents are Chemical shifts (ppm) also in accord with the postulate that 3e forms more stable and Mole fraction more ordered complexes with polar solvents than does 3a. In 8 THP 8 2-4 C-4 [8c-41 n3e other words, we contend that the changes in free energy with changes in solvent presented in Table 5 have their origins in Additivea at 308 K the differences between the polar interactions of the solvent None 18.790 0 18.79 0.18 molecules with the two conformers. To reiterate, the polar DMSO 18.878 0.027 18.91 0.21 interactions are considered to be stronger for an equatorial DTBPC 18.818 0.027 18.85 0.20 conformer because the absence of an endo-anomeric effect DTBPC-DMSO (1:l) 18.928 0.055 18.98 0.22 renders further charge delocalization energetically more facile. PhOH 0.578 19.34 18.762 0.29 By affecting the charge delocalizations within the conformers, PhOH-DMSO (1:l) 18.95 0.163 19.12 0.25 polar and hydrogen-bonding solvents must influence the magniAt 268 K tudes of the endo- and exo-anomeric effects, but these intramolecular effects cannot be separated from the forces of None 18.693 -0.029 18.71 0.17 association. DMSO 19.796 0.039 18.84 0.19 DTBPC 18.683 These studies have shown that stereoelectronic influences 0.049 18.73 0.17 DTBPC-DMSO(1:l) 18.844 0.054 18.90 0.20 on molecular structure are affected by interactions with polar PhOH 18.639 0.29 0.701 19.34 molecules. Effects on the exo-anomeric effect present in PPhOH-DMSO (1:l) 18.87 0.138 19.01 0.23 glycopyranosides wherein the aglycon is in equatorial orientation can be expected to be especially large. In the case of the a "DMSO = dimethylsulfoxide; DTBPC = 2,6-di-rert-butyl-4-methylphenol; anomers, the solvation influences the strengths of both the endoPhOH = phenol. All additions rendered the solution 0.8 M in the additive. and exo-anomeric effects. However, there can be n o doubt that hydration of the ring oxygen will strengthen the exo-anomeric DMSO increased n3, to only 0.21. The addition of both these effect for both the anomers and thereby increase its contribution compounds should increase rile to a lesser extent since the to the conformational rigidity of all glycosides. In this regard, DMSO would compete with 3 for bonding with the phenol. This the contributions used in HSEA calculations (12) that refer to was the case (Table 4). Similar experiments were conducted an isolated molecule of dimethoxymethane are likely minimum using the highly hindered 2,6-di-tert-butyl-4-methylphenol, to values. It is of interest to note that simple rotation about a strengthen the evidence that hydrogen bonding is responsible glycosidic bond must weaken the exo-anomeric effect and for the increase in the population of the equatorial conformer thereby activate the anomeric carbon toward nucleophilic attact. that occurred on adding phenol. As seen in Table 4, addition Likely, such rotations occur in the course of the formation of of this hindered phenol (DTBPC) had little influence on the activated complexes with enzymes that catalyze transformaconformational equilibrium whether or not DMSO was present. tions involving the replacement of the aglycons of glycosides. This is as expected, since the OH group of DTBPC is too hindered to form intermolecular hydrogen bonds. Experimental As was discussed in the Introduction, the change in enthalpy General procedures for the conformational equilibrium 3a i2 3e should be positive Except for D20, all solvents were purified and stored over 3.A and about 1 kcal/mol. Table 5 reports the effects of changes in molecular sieves (BDH, Toronto). temperature and solvent on the equilibrium 3a?z 3e as assessed 'The 2-alkoxytetrahydropyrans (3, 11, 12,13) were prepared, in the by 13Cnmr. The effects of temperature change were then used to usual manner, by hydrogen chloride catalyzed addition of the alcohol to estimate the changes in enthalpy and entropy. As seen in Table 2,3-dihydro-4H-pyran; 12: bp 69"C, 17 Torr (1 Torr = 133.3 Pa); 5, with the nonpolar and weakly associating solvent CC14-C6D6 13: bp 126-127"C, 20 Torr. Deacetylation of 13 provided the (lo%), the change in enthalpy (0.8 kcal/mol) is in fact nearly 2-(2-hydroxyethoxy)tetrahydropyran (15) (bp 73-77"C, 3 Torr). Tri1 kcal/mol and the conformational change involves a very small methylsilylation of the 2-hydroxy tetrahydropyran (16) (36) readily decrease in entropy. For reasons to be discussed below with afforded 14. reference to more polar solvents, it is apparent that C6F6 The addition of either hexadeuteriobenzene or hexafluorobenzene (10%) to carbon tetrachloride was made for field frequency lock associates more strongly with 3 than does C6D6, and this is purposes. reasonable because of the presence of much more polar bonds Unless otherwise stated, the various solutions of the 2-substituted in C6F6. The change in enthalpy with CDC13 as solvent was tetrahydropyrans were 0.1 M. These solutions were 0.05 M in tetra-0.2 kcal/mol, a value that is near the zero value observed hydropyran, which served as the internal standard. by Booth et at. (24) using a mixture of CDC13 and CFC13 All the I3C spectra recorded at 90.56 MHz in the FT mode as solvent. Thus, in CDC13, the change in free energy used a Bruker WM 360 spectrometer interfaced with an Aspect-2000 (0.48 kcal/mol) is mainly determined by the decrease in computer for data accumulation and processing. The temperature was entropy. This is not surprising, since CDC13must form (through regulated with a VT-1000 temperature controller. Except for the hydrogen bonding) much better defined (ordered) complexes studies involving 4, the same spectrometer was used to measure the with 3 by associating with the acetal oxygens than it can through 'H nrnr spectra (360.13 MHz). self-association by way of extremely weak intermolecular 2-Methoxy-4,4,5,5-tetradeuterio-tetrahydropyran (6) hydrogen bonds. Furthermore, as was to be expected from the The compound was prepared as previously reported (11) except foregoing discussions, the interaction of CDC13 with 3e should that pent-3-yne-5-01-1-a1 dimethyl acetal was acetylated prior to the be stronger than with 3a. Therefore CDC13 should decrease the deuteration. The product was deacetylated (sodium methoxide enthalpy for 3e to a greater extent than for 3a and thereby, as methanol) prior to treatment with a solution of methanol saturated with was observed, lead to a smaller change in enthalpy in this hydrogen chloride in dichloromethane for 15 min at O°C. The product TABLE4. The effect of intermolecular hydrogen bonding on the conformational equilibrium of 0.1 M 2-MeO-THP (3) in CC14-C6D6

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

219

220

CAN. J. CHEM. VOL. 6 5 , 1987

TABLE5. Approximate thermodynamic parameters for the conformational equilibria of 2-MeO-THP (3) in different solvents Chemical shifts (pprn)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

Solvent

Temperature (K)

Observed

-

"Absolute"

6c-4

[SC-~]

nc

AGO (kcal mol-')

Kc

AHO

(kcal mol-I)

As0 (cal K-' mol-')

TABLE6. Determination of the conformational equilibria at 308 K of 6 in a variety of solvents from the coupling of H-6c and H-6r with deuterium atoms at C-5 Half-height widths Chemical shifts (ppm) Solventn

H-6c

H-6r

+

J H , ~ JH,H~

H-6c

H-6t

JH.H~

H-6c

H-6t

JH.D~

H-6c

H-6t

nkn

"Mole fraction 6e = half height width H-6c divided by the sum of the half heights of H-6c and H-6t (see Table 1). bWithout decoupling. 'With deuterium decoupling. dBy subtracting the values for J with and without deuterium coupling.

(73% yield) was isolated by distillation and contained about 10% dichloromethane. The proton and deuterium decoupled I3cspectrum was recorded at 100.61 MHz. The 5020 scans were accumulated at 308 K using a 0.5-mL sample containing 10 pL of 2-methoxytetrahydropyran (3) and 40 pL of the 4,4,5,5-tetradeuterated derivative (6) with 10% hexafluorobenzene in carbon tetrachloride as solvent. To assign the chemical shifts, a spectrum was recorded under the same conditions but in the absence of 3. The gc-ms using a 30-m capillary silicon-coated (DB-1) column (temperature increase: g°C/min from 60 to 200°C) produced the following relative intensities for the various tetrahydropyranyl ions in the range 86 5 m / e 5 91: 9, 25, 70, 100, 43, 9. Thus the product was a mixture of several deuterated derivatives of 3 with the main components being tri (27%), tetra (37%), and penta (18%) species. Nevertheless, the material proved useful for the planned application. In this regard, the ' H nrnr spectra in the various solvents (Table 2) all showed a major component that provided, on deuterium decoupling, a

well-defined ABX pattern for H-3c, H-3t, and H-2 and an AB pattern for H-6c, H-6t. An examination of the 13Cnmr spectrumof the mixture under conditions of both hydrogen and deuterium decoupling showed that these signals arose from the 4,4,5-tri- and 4,4,5,5-tetradeuterio compounds (about 50% abundance). The 13cnmr spectrum confirmed the presence of the 4,5,5- and 4,4,5-trideuterio and the 4 , 4 3 3 tetradeuterio compounds for the following reason. It is known (37) that the effect of deuteration on a 13Catom in cyclohexane is an upfield shift of 0.418 pprn when the separation is by one bond, 0.104 pprn by two bonds, and 0.025 pprn by three bonds. Since the chemical shift for C-3 of MeO-THP (3) under the conditions used was 30.519 ppm, it could be expected that the mixture would produce signals for C-3 at 30.365, 30.286, and 30.261 ppm for the 4,5,5-,4,4,5-, and4,4,5,5-deuterated compounds, respectively. In fact, signals occurred at 30.366,30.285, and 30.260 ppm. Low intensity signals at 30.013 and 29.900 pprn suggest that some exchange of H-3 atoms by deuterium had occurred. Since the deuteration was conducted prior to ring formation, the

PRALY AND LEMIEUX

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

TABLE7. Effects of solvent and temperature on the I3C chemical shifts of TMS and C-4 of THP

1

1

I /

1

I 1 I

I

221

and 10. The spectra of course displayed much weaker signals in close proximity to these main patterns, which are attributable to the other deuterated derivatives of 3 present in the mixture. Integration of the spectra and assuming that the methoxy group contained only natural Observed chemical abundance deuterium required very little, if any, deuteration of C-2 or shifts (ppm) ~ c - 4 ~ ~ C-6. " As mentioned above, some deuteration of C-3 had occurred. Temperature For an nOe study of 6 at 360.13 MHz, specially treated (40) 5-rnm Solvent (K) TMS C-4 THP (ppm) glass tubes were used. Argon was bubbled throughout the solutions for 5 min, then the tubes were immediately sealed. For these experiments, CCI4-C6D6 (10%) 348 -0.195 23.585 99.96% enriched deuterium oxide with low paramagnetic impurity 308 0.0 23.618 content was used. The data were obtained at 308 K in the difference 288 0.119 23.642 mode from at least 950 accumulated scans. 268 0.209 23.639 The signals for the hydrogens at C-6 of 6 were readily assigned since saturation of H-2, while observing the signals for the hydrogens at C-6, CC14-C6F6 (10%) 308 0.002 23.840 showed that only one of these, the higher field signal, was enhanced 268 0.003 23.684 and, therefore, the signal for the hydrogen (H-6c), which is in CDC13 328 -0.243 23.402 cis-relationship to H-2. The signal for the methyl group was also 308 -0.134 23.403 enhanced. For the two conformers, the distances between the average 288 -0.033 23.401 positions of the methyl group hydrogens and H-2 are very nearly the 268 0.041 23.374 same. Therefore, the nOe's observed for the methyl group signals CD3CN 348 4.134 28.479 should be nearly equilibrium independent. Consequently, the ratio R 308 4.181 23.470 for the nOe's observed for H-6c and the methyl group under a given set 288 4.208 28.469 of experimental conditions should be proportional to the point of 268 4.230 28.451 equilibrium and should increase as the population of 3e increases. The R values in three different solvents are reported in the footnote to 25.553 D20 348 Table 1. That these values correspond to the mole fractions of 3e is 25.020 308 surely coincidental. This finding reinforces the contention that 6 exists 24.669 288 essentially only in conformers 6a and 6e. 278 24.534 Further support for the equilibria reported in Table 1 was provided by the coupling of H-6c and H-6t with deuterium atoms at C-5 in the minus 23.618 ppm. "Observed';6: deuterated 2-MeO-THP preparation discussed above with reference to Table 1. As seen in Table 6, the decoupling of the deuterium atoms trideuterated compounds must be present in the mixture as equal reduced the half-band widths observed for H-6c and H-6t and thereby amounts of their diastereoisomeric forms; namely, 7 and 8 for the provided a measure of the relative strengths of the deuterium couplings. 4,4,5- and 9 and 10 for the 4,5,5-deuterated compounds. It is expected that the coupling of the axial H-6t of the 6a conformer with the deuteriums at C-5 will be the same as that for the axial H-6c of conformer 6e, since their orientations relative to the deuteriums must be very similar, if not the same. Similarly, H-6c of 6a is in an environment also present for H-6t of 6e. Therefore the ratio of the change in the coupling with deuterium of one of the H-6's to the sum of the changes involving both H-6t and H-6c should provide a measure of the mole fraction of one of the two conformers. In fact, as seen in Table 6, the deuterium coupling with H-6c increased on increasing the polarity of the solvent in accord with an increasing abundance of 6e. This follows, since in conformer 6e, H-6c is anti-periplanar and syn-clinal to deuteriums at C-5 but is syn-clinal to both the deuteriums at C-5 in the case of 6a. As seen in Table 1, the conformational equilibria estimated in this manner proved to be in fine agreement with those found by the other procedures. Conformational equilibria by ' H nuclear magnetic resonance The proton spectra of 2-methoxy-4,4,5,5-tetradeuteriotetrahydro- Conformational equilibria by I3cnuclear magnetic resonance pyran, under deuterium decoupling (61.42 MHz, 30 db), were recordThe frequency corresponding to tetramethylsilane (TMS) in CC14ed at 400.13 MHz using a 2400-Hz spectral window and 32K data C6D6 (10%) at 308 K was taken as the ultimate external reference points (digital resolution after transformation: 0.15 Hz/pt). Since the chemical shift. To be sure that no appreciable drift in magnetic field had occurred, the position of this reference signal at the prevailing deuterium signal was useless for lock purposes and in the absence of fluorine in the samples, one single scan was recorded after a 90" pulse. temperature was checked by repeating selected experiments at some Prior to Fourier transformation, resolution enhancement was achieved later date. The chemical shift for C - 4 of tetrahydropyran (THP) was through a Lorentz-Gauss transformation (0.25 5 Gaussian broadening then used as internal standard to adjust the observed chemical shifts for solvent and temperature effects. This was expected to be useful since 5 0.4, -0.7 5 line broadening 5 -0.55), then the spin parameters were optimized through iterative calculations using the LAOCOONthe environments for C-4 in conformers 3a and 3e are very similar. The 111-like (38) PANIC program (39) for the above mentioned Aspect choice of the signal for C-4 in tetrahydropyran as a reference signal was also attractive because Booth et al. (24) had shown that there exists an computer, resulting in a 20.02 Hz probable error for the coupling important (4.59 ppm) chemical shift difference at 143 K for C-4 in constants assigned to H-2, H-3c, H-3t, H-6c, and H-6t. these conformers. For these reasons, the chemical shifts for C-4 of As was mentioned above, the ' H nmr spectra of the mixture THP, relative to the external TMS signal, are reported in Table 7. comprising mainly (37%) 6 but about equal amounts of 7-10 (each These shift differences (AC-4THP)are then used in Table 8 to adjust the about 7% of the mixture) produced, on deuterium decoupling, readily observed values for C-4 of 2-MeO-THP (3) to the so-called "absolute" distinguishable ABX patterns assignable to compounds 6 , 7 , and 8 and an A'B' pattern assignable to H-6c and H-6t of compounds 6, 9, chemical shifts, i.e., chemical shifts that are expected to be indepen-

CAN. J. CHEM. VOL. 65. 1987

TABLE8. Data for the calculation of "absolute" chemical shifts for C-4 of 3 a and 3e Chemical shifts (ppm)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

Solvent

Expt. No.

Temperature (K)

Mole fraction" n3e

Observed "~bsolute" Sc-4 [Sc.,]

CC14-C6D6 (10%)

1 2 3

308 278 268

0.20 (0.16) 0.15

18.818 18.679

18.82 (18.70) 18.66

CC14-C6F6 ( 10%)

4 5 6

308 27 8 268

0.20 (0.18) 0.17

19.042

CDC13

7 8 9

308 278 268

0.33 (0.33) 0.33

CD3CN

10 11 12

308 278 268

0.37 (0.36) 0.35

18.774 19.280 19.288 24.588 24.525

18.82 (18.74) 18.71 19.50 (19.52) 19.53 19.74 (19.70) 19.69

D20

13 14

308 278

0.48 0.47

21.628 21.162

20.23 20.25

-

"From Table 3 with the values at 278 K in parentheses obtained by interpolation. bBy substitution of the A:!! values in Table 4 from the observed chemical shift for C-4 of 3. The values in parentheses were obtained by interpolation.

TABLE9. Calculation of "absolute" chemical shifts for C-4 of 3 in the axial (3a) and equatorial (3e) conformers "Absolute" chemical shifts (ppm) Combinationsa

[6~-413e

[6~-4I3a

At308K 1- 7 1-10 1-13 4- 7 4-10 4-13 7-13 10-13

23.01 23.15 22.85 23.01 23.15 22.85 22.76 22.55 22.92 -

17.77 17.74 17.81 17.77 17.74 17.81 17.89 18.09 17.83 -

22.75 22.90 22.90 23.00 23.11 23.01 23.01 22.90 22.95 -

17.93 17.90 17.90 17.80 17.78 17.80 17.80 17.90 17.85 -

22.77 23.04 22.96 23.23 -

17.93 17.89 17.84 17.78 -

23.00 -

17.86 -

Average: At278 K 2- 8 2-11 2-14 5- 8 5-1 1 5-14 8-14 11-14 Average: At268K 3 - 9 3-12 6-9 6-12 Average: Overallaverage

22.96?0.16

17.85?0.09

[8~-413e- [6~-413e (PP~)

5.09

dent of solvent and temperature. On this basis, these chemical shifts were used in conjunction with the conformational equilibria reported in Table 3 to estimate "absolute" chemical shifts [ 6 ] for C-4 in the conformers 3 a and 3e. This was accomplished by solving the sets of simultaneous equations presented in Table 9. The data in Table 9 indicate that effects of changes in temperature on chemical shifts were not detected. Consequently, the simple expression was used to estimate the conformational equilibria, not only for compound 3, but also for other 2-substituted tetrahydropyrans, since it was expected that changes at C-2 would not significantly affect the chemical shift of C-4. Following this procedure, the only requirement is that the observed chemical shift for C-4 be corrected for the general solvent and temperature effects using the A values for C-4 of tetrahydropyran as provided in Table 7. Application of the procedure is justified since it well reproduces the conformational equilibria measured by way of ' H nmr. Although its accuracy cannot be predicted, the procedure is certainly not misleading.

Acknowledgement

5.10

5.14 5.11?0.25

"Using the data in Table 5 and the expression, [8C.4] = n3e[8c4]3e + n3a[sc-413a.

T h e authors are particularly indebted to D r . Ole Hindsgaul for his supervision of the w o r k done on the 360-MHz 'H nmr spectrometer, but also deeply thankful for the guidance b y Dr. T. Nakashima and G . Bigam of the nrnr-spectral service laboratory of the department. T h e research was supported b y the Natural Sciences and Engineering Research Council of Canada (Grant A-172 to R.U.L.) a n d by the Alberta Heritage Foundation for Medical Research (grant to R.U.L.). Abstr. Pap. Am. Chem. Soc. 135, 5E (1959). 1. R. U. LEMIEUX. and P. CHU. Abstr. Pap. Am. Chem. Soc. 133, 2. R. U. LEMIEUX 31N (1958). 3. R. U. LEMIEUX, R. K. KULLNIG, H. J. BERNSTEIN, and W. G. SCHNEIDER. J. Am. Chem. Soc. 80, 6098 (1958). 4. A. J. KIRBY.The anomeric effect and related stereoelectronic effects at oxygen. In Reactivity and structure concepts in organic chemistry. Vol. 15. Springer Verlag, New York. 1983.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/16/17 For personal use only.

PRALY AND LEMIEUX

Stereoelectronic effects in organic chemis5. P. DESLONGCHAMPS. try. Pergamon Press, New York. 1983; (a) pp. 4-53; (b) p. 31. Molecular rearrangements. Part Two. Edited by 6. R. U. LEMIEUX. P. de Mayo. Interscience, New York. 1964. p. 709. 7. C. ALTONA.Ph.D. Thesis, University of Leiden, 1964. and D. J. MITCHELL. Carbohydr. 8. S. WOLFE,M.-H. WHANGBO, Res. 69, 1 (1979). J. A. POPLE,J. S. BINKLEY, and S. VISHVESH9. G. A. JEFFREY, WARA. J. Am. Chem. Soc. 100, 373 (1978). S. KOTO,and 10. R. U. LEMIEUX,K. BOCK,L. T. J. DELBAERE, V. S . RAO.Can. J. Chem. 58, 631 (1980). 1 1 . R. U. LEMIEUX,A. A. PAVIA,J. C. MARTIN,and K. A. WATANABE. Can. J. Chem. 47, 4427 (1969). N , U. LEMIEUX,K. BOCK,and B. MEYER. 12. H. T H ~ G E R S ER. Can. J. Chem. 6 0 , 4 4 (1982). 13. R. U . LEMIEUX,S. KOTO, and D. VOISIN.ACS Symp. Ser. No. 87. 1979. Chapt. 2. p. 17. and T. KOZAR.J. Am. Chem. Soc. 104, 6929 14. I. TVAROSKA (1980). , SCHLIEFER, and E. TARTAKOVSKY. NOUV.J. Chim. 15. B. F u c ~ sL. 8, 275 (1984). 16. A. J. BRIGGS,R. GLENN,P. G. JONES,A. J. KIRBY,and P. RAMASWAMY. J. Am. Chem. Soc. 106, 6200 (1984). and R. U. LEMIEUX. 17. R. BALL,A. P. VENOT,0 . HINDSGAUL, Manuscript in preparation. 18. G. M. LIPKIND,V. E. VEROVSKY, and N. K. KOCHETKOV. Bioorgan. Khim. 9, 1269 (1983);Carbohydr. Res. 133, 1 (1984). A. P. VENOT,U. SPOHR,P. BIRD,G. MANDAL, 19. R. U. LEMIEUX, N. MORISHIMA, 0 . HINDSGAUL, and D. R. BUNDLE.Can. J. Chem. 63, 2664 (1985), and references therein. 20. R. U . LEMIEUX and A. A. PAVIA.Can. J. Chem. 47,4441 (1969). 21. I. TVAROSKA and T. KOZAR.Int. J. Quantum Chem. 23, 765 (1983). 22. R. U. LEMIEUX.Proceedings VIIth International Symposium on

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

223

Medicinal Chemistry. Vol. 1 . Swedish Pharmaceutical Press, Stockholm. 1985. p. 329. H. BOOTHand K. A. KHEDHAIR. J. Chem. Soc. Chem. Comrnun. 467 ( 1 985). H. BOOTH,T. B. GRINDLEY, and A. K. KHEDHAIR. J. Chem. Soc. Chem. Comrnun. 1047 (1982). A. J. DE HOOG.Org. Magn. Reson. 6, 233 (1974). B. M. PINTOand S. WOLFE.Tetrahedron Lett. 23, 3687 (1982). R. W. FRANCK. Tetrahedron, 39, 3251 (1983). E. L. ELIELand M. H. GIANNI.Tetrahedron Lett. 97 (1962). G. A. PIERSONand 0 . A. RUNQUIST. J. Org. Chem. 33, 2572 (1968). B. F u c ~ s ,A. ELLENCWEIG, E. TARTAKOVSKY, and P. APED. Angew. Chem. 98, 289 (1986). G. A. JEFFREY and J. MITRA.Acta Crystallogr. Sect. B, 34a, 469 (1983). A. EL-KAFRAWY and R. PERRAUD. C.R. Acad. Sci. Paris, 280, 1219 (1975). B. JOHNSEN and P. E. S ~ R E N S EActaChem. N. Scand. Ser. A, 33, 241 (1979). S. J. ANGYAL,V. A . PICKLES,and R. AHLUWALIA. Carbohydr. Res. 3, 300 (1967). J. P. PRALY,G. DESCOTES, R. FAURE,and H. LOISELEUR. Cryst. Struct. Cornmun. 11, 1323 (1982). L. E. SCHNIEPP and H. H. GELLER.J. Am. Chem. Soc. 68, 1646 (1946). R. AYDINand H. GUNTHER.J. Am. Chem. Soc. 103, 1301 (1981). S. COSTELLANO and A. A. BOTHNER-BY. J. Chem. Phys. 41, 3863 (1964). PANIC Program: Copyright, Bruker Spectrospin AG, Switzerland. Y. THERIAULT, Y. MORIO,and G. KOTOVYCH. Can. J. Biochem. 57, 165 (1979).